818167

research-article2018
AUT0010.1177/1362361318818167AutismAu-Yeung et al.

Original Article

Autism

Experience of mental health diagnosis 1­–11

© The Author(s) 2018
Article reuse guidelines:
and perceived misdiagnosis in autistic, sagepub.com/journals-permissions
DOI: 10.1177/1362361318818167
https://doi.org/10.1177/1362361318818167

possibly autistic and non-autistic adults journals.sagepub.com/home/aut

Sheena K Au-Yeung1 , Louise Bradley1, Ashley E Robertson1 ,

Rebecca Shaw1,2, Simon Baron-Cohen3,4 and Sarah Cassidy1,3,5

Abstract
Previous research shows that autistic people have high levels of co-occurring mental health conditions. Yet, a number
of case reports have revealed that mental health conditions are often misdiagnosed in autistic individuals. A total of
420 adults who identified as autistic, possibly autistic or non-autistic completed an online survey consisting of questions
regarding mental health diagnoses they received, whether they agreed with those diagnoses and if not why. Autistic and
possibly autistic participants were more likely to report receiving mental health diagnoses compared to non-autistic
participants, but were less likely to agree with those diagnoses. Thematic analysis revealed the participants’ main reasons
for disagreement were that (1) they felt their autism characteristics were being confused with mental health conditions
by healthcare professionals and (2) they perceived their own mental health difficulties to be resultant of ASC. Participants
attributed these to the clinical barriers they experienced, including healthcare professionals’ lack of autism awareness
and lack of communication, which in turn prevented them from receiving appropriate support. This study highlights
the need for autism awareness training for healthcare professionals and the need to develop tools and interventions to
accurately diagnose and effectively treat mental health conditions in autistic individuals.

Keywords
adults, autism spectrum disorders, diagnosis, mixed methods, prevalence, psychiatric comorbidity, qualitative research

Introduction during their lifetime, with mood and anxiety disorders

Autism spectrum condition (ASC)1 is classified as a neu-
rodevelopmental condition, characterised by persistent
difficulties in social communication and interaction, as
well as restricted, repetitive patterns of behaviour, interests
and activities (American Psychiatric Association (APA),
2013). These symptoms present during early development
and cause a range of strengths and difficulties in multiple
domains. Recent studies suggest that autistic difficulties
could increase the likelihood of being exposed to risk fac-
tors for developing mental health conditions. For example,
higher autistic traits were associated with self-reported
loneliness in autistic adults (Hedley et al., 2018) and the
experience of social disconnectedness, and feeling a bur-
den on others in the general population (Pelton and
Cassidy, 2017); these factors in turn increased the risk of
both depression and suicidality. Indeed, Lever and Geurts
(2016) established that 79% of autistic adults met the cri-
teria for some form of psychiatric co-occurring condition
being the most common, and at a significantly greater rate
compared to non-autistic adults.
Despite increased risk factors for, and incidence of,
mental health conditions in autistic individuals, conditions
such as depression may still be under-diagnosed given the
lack of validated assessment tools able to accurately detect
the unique presentation of mental health difficulties in this
group (Cassidy et al., 2018a, 2018b). Although ASC is not
1Coventry University, UK
2Coventry and Warwickshire Partnership NHS Trust, UK
3University of Cambridge, UK
4Cambridge Lifespan Asperger Syndrome Service (CLASS), UK
5University of Nottingham, UK
Corresponding author:
Sheena K Au-Yeung, Centre for Innovative Research Across the Life
Course, Coventry University, Richard Crossman 4th Floor, Jordan
Well, Coventry CV1 5FB, UK.
Email: Sheena.Au-Yeung@coventry.ac.uk
2 Autism 00(0)
a mental health condition, the characteristics of ASC can
overlap with many indicators of mental health conditions;
for example, social withdrawal, appetite and sleep distur-
bance are common characteristics in both ASC and depres-
sion (Stewart et al., 2006). This means that an established
diagnosis of ASC could overshadow the presence of men-
tal health symptoms, potentially leaving co-occurring
mental health conditions unrecognised (Matson and
Williams, 2013). The consequences of failing to accurately
diagnose co-occurring mental health conditions are poten-
tially devastating and increase the risk of contemplating,
planning (Cassidy et al., 2014) and dying by suicide
(Hirvikoski et al., 2016).
It is also possible that symptom overlap could lead to
overdiagnosis and misdiagnosis of mental health condi-
tions in autistic individuals (Lai and Baron-Cohen, 2015).
In a case series, Dosseter (2007) re-evaluated four young
clients, aged 8–16 years, who had been originally diag-
nosed with early-onset psychosis. The individuals reported
unusual experiences, exhibited bizarre behaviours and had
preoccupations with objects and imaginary people or
worlds. However, when the clients’ impairments in social
communication and reciprocity were taken into considera-
tion, it was concluded that clients’ presentation could be
better explained by ASC, with the preoccupations
accounted for under ‘restricted, repetitive patterns of
behaviour, interest, and activities’ (APA, 2013). Other case
reports (Luciano et al., 2014; Van Schalkwyk et al., 2015)
also indicated that mental health professionals’ failure to
collect information about developmental history, and lack
of understanding about ASC presentation, contributed to
mental health misdiagnoses in both individuals with an
existing autism diagnosis and those who obtained a diag-
nosis during the studies. Misdiagnosis of ASC as a mental
health condition could lead to unnecessary treatment (e.g.
medication), which could preclude the possibility of hav-
ing other, more relevant, needs met.
While previous case studies (Dosseter, 2007; Luciano
et al., 2014; Van Schalkwyk et al., 2015) provide impor-
tant clues to the contributing factors for mental health
misdiagnosis in autistic individuals from the perspective
of professionals specialising in ASC, there is currently
little known about autistic individuals’ subjective experi-
ence of mental health diagnosis and perceived misdiagno-
sis. Therefore, this article aims to address the following
research questions: (1) Are autistic individuals more
likely to report receiving mental health diagnosis(es) and
if so, what kind of diagnoses were they more likely to
receive? (2) Are autistic individuals more likely to disa-
gree with these diagnoses? and (3) What were the reasons
behind any disagreement?
Three participant groups were included, comprising
individuals who were identified as autistic, non-autistic or
possibly autistic. Inclusion of the non-autistic group
allowed us to examine whether mental health diagnoses
and perceived misdiagnoses were more common in autis-
tic individuals compared to the general population. We
included a group of possibly autistic people who reported
suspecting that they may have an undiagnosed ASC. The
reasons for this are that yet-to-be-diagnosed autistic indi-
viduals may be particularly vulnerable to misdiagnosis of
mental health conditions (Dosseter, 2007; Luciano et al.,
2014; Van Schalkwyk et al., 2015) and also that autistic
individuals often play an important role in the recognition
of their ASC. Jones et al. (2014) looked at the experience
of those who had received an ASC diagnosis and found
that 44.5% of their respondents reported raising the possi-
bility of ASC themselves, whereas only 4.1% reported that
a professional raised the possibility. It is also well known
that certain characteristics affect the likelihood of receiv-
ing an autism diagnosis, for example, females are diag-
nosed less frequently than males (Brugha et al., 2011).
This may be due to differences in presentation, such as
better nonverbal communication skills and the ability to
camouflage other diagnostic features (Rynkiewicz et al.,
2016). Seeking a diagnosis in adulthood may be problem-
atic due to the development of compensatory strategies
and a possible lack of informant (e.g. parent) for develop-
mental history (Lai and Baron-Cohen, 2015). For these
reasons, it is important for the experiences of this previ-
ously under-researched group, namely, possibly autistic
adults, to be heard.
Method
Materials
Survey development. Research priorities for the autism
community can be very different from those of the
researchers (Pellicano et al., 2014). Therefore, it is impor-
tant that autistic people participate in setting priorities to
ensure that the research conducted is useful to those most
affected by it (Fletcher-Watson et al., 2018). As such, an
online survey exploring various topics including mental
health diagnoses, self-injury and suicidality was devel-
oped in partnership with a steering group of eight autistic
adults (six females, two males) through a series of six
focus groups (see Cassidy et al., 2018c, for details of the
full data set). At the start of the study, the steering group
discussed what their role should be and decided it would
be to inform the study conception, design, recruitment and
dissemination of results. The first three focus groups
focused on developing the questions in the online survey
about a number of topics identified from a systematic
review of the literature thought to be important contribu-
tors to mental health and suicidality in autism. The focus
group fed back on the relevance of these proposed topics,
whether any important topics were missing and their own
experiences in relation to the topics identified. The
researchers then developed a survey with questions that
Au-Yeung et al. 3
would allow participants’ experiences of these topics to be
captured. Once the survey had been designed, the steering
group provided feedback on three drafts until agreement
was reached that the questions were comprehensive, rele-
vant and clear. The data collected in relation to mental
health diagnoses were analysed and are reported in this
article. Data relating to other topics collected from the sur-
vey are reported elsewhere (see Camm-Crosbie et al.,
2018).
Mental health diagnosis questions. Participants answered
closed- and open-ended questions on their mental health
diagnoses. First, they were asked ‘Have you ever been
diagnosed with a mental health condition(s) or other
condition(s) by a trained clinician?’ to which they
responded with either ‘Yes’ or ‘No’. They were subse-
quently asked to select the specific mental health or other
condition(s) including ‘Depression, Anxiety, Obsessive
Compulsive Disorder (OCD), Bipolar, Personality Disor-
der, Schizophrenia, Anorexia, Bulimia and Other’. If par-
ticipants selected ‘Personality Disorder’ or ‘Other’ from
the list, they were given a free text box to specify the sub-
type of Personality Disorder or other mental health condi-
tions they were diagnosed with. Participants were then
asked, ‘Do you agree with your diagnosis/es’, with the fol-
lowing response options: ‘Yes’, ‘No’ or ‘Not entirely’. If
participants selected ‘No’ or ‘Not entirely’, they were sub-
sequently asked, ‘Please tell us what mental health diagno-
sis/es you disagree with, and why’ and indicated their
response using a free text box.
These diagnoses and order were chosen due to a combi-
nation of standard questions our group regularly uses in
previously published research and also feedback from the
steering group regarding clarity. For example, participants
noted that it was easier to have staged questions starting
wide, and then narrowing (Personality Disorder subtype),
to keep the survey questions relevant throughout.
Demographic questions.  As part of the survey, participants
were asked to provide information such as age, gender,
autism diagnosis and age of diagnosis. Participants were
divided into three groups for analysis based on their autism
diagnostic status. Participants in the non-autistic group
reported no formal diagnosis of ASC and no suspicion that
they have ASC. Participants in the autistic group reported
receiving a formal diagnosis of ASC from a qualified clini-
cian. Participants in the possibly autistic group reported
that they thought they may have ASC and either (1) were
awaiting assessment from a qualified clinician or (2) had
not yet sought an assessment.
Autism Spectrum Quotient.  The Autism Spectrum Quotient
(AQ; Baron-Cohen et al., 2001) is a measure of autistic
traits with scores ranging from 0 to 50. Greater scores indi-
cate higher levels of autistic traits. Participants rated
themselves on 50 statements with the options ‘Definitely
Agree’, ‘Slightly Agree’, ‘Slightly Disagree’ and ‘Defi-
nitely Disagree’. Woodbury-Smith et al. (2005) recom-
mended a cut-off of 26 for screening in clinical practice
and a cut-off of 32 for the general population, to limit false
positives. We followed Hoekstra et al.’s (2007) methodol-
ogy to manage missing items. If five or fewer items were
left blank, the following equation was used to correct for
the number of missing items: total subscale score + (mean
subscale item score × number of missing items in sub-
scale). The corrected subscale scores were summed to get
the total AQ score. Missing items accounted for 0.19% of
the analysed data. If a participant left more than five items
blank, their data were discarded. Two participants’ data
were excluded based on this criterion.
Participants
Participants were recruited through charities, Cambridge
Autism Research Database (CARD), Cambridge
Psychology, online advertisement, MHAutism newsletter
and MHAutism webpage (http://mhautism.coventry.
ac.uk). Participants were invited to take part in an online
survey exploring experiences of mental health problems,
self-injury and suicidality and could take part regardless of
prior experience of these difficulties. Data from 420 adults
(18–67 years) who completed the ‘mental health diagno-
ses’ section of the online survey were analysed. Data from
respondents who did not report their age (n = 52), gender
(n = 3) or diagnostic status (n = 0) were excluded. The
autistic group was significantly younger than the non-
autistic group (p = 0.013), and there was a significantly
greater proportion of males and smaller proportion of
females compared to both the possible autistic and non-
autistic groups (ps > 0.05). Welch’s analysis of variance
(ANOVA) revealed that there were significant between-
group differences in AQ scores, F(2, 189) = 281, p < 0.001.
Autistic and possibly autistic participants scored signifi-
cantly higher on the AQ than non-autistic participants
(ps < 0.001). Autistic participants also scored significantly
higher than possibly autistic participants on the AQ
(p = 0.005). Participants’ characteristics are reported in
Table 1.
Procedure
The study was approved by Coventry University
Psychology Ethics Committee, the autism steering group
who fed back on the questionnaire and the scientific
advisory group at the Autism Research Centre, University
of Cambridge. Participants were invited to complete
the survey on Qualtrics (https://www.qualtrics.com/uk/).
Participants read the participant information sheet and
indicated informed consent to participate via an online
form. Participants were fully briefed about the nature of
4 Autism 00(0)
Table 1.  Participants’ characteristics.
Autistic (n = 208)
Mean age (SD), range 38.6 (11.4), 18–67
Number of males 72 (34.6%)
Number of females 136 (65.4%)
Mean AQ (SD) 38.6 (6.8)
Mean age of diagnosis (SD), range 34.5 (12.2), 4–59
SD: standard deviation.
Table 2.  Number of participants who reported diagnosis of mental health condition or other conditions.
‘Have you ever been diagnosed with a mental health
condition(s) or other condition(s) by a trained clinician?’
Yes
No
the research, that they could skip sections and/or ques-
tions that they felt uncomfortable with, and were provided
with information about relevant support services before
and after participation.
Analysis
Quantitative analysis. Statistical analyses were conducted
using SPSS version 24. Chi-square tests for homogeneity
were computed to test for between-group differences in the
proportions of (1) individuals reported receiving mental
health and other diagnoses, (2) specific mental health diag-
noses received and (3) agreement with mental health diag-
noses. If a chi-square was significant, post-hoc z tests were
then computed to further examine pairwise differences in
group proportions with automated adjusted p-values (Bon-
ferroni method). Fisher’s exact tests were used instead if
any of the expected cell count were less than 5. Cramer’s
V was calculated for effect size (small effect = 0.1, medium
effect = 0.3, large effect = 0.5; Cohen, 1988).
Qualitative analysis.  Thematic analysis was conducted for
participants’ responses to the question, ‘Please tell us what
mental health diagnosis/es you disagree with, and why’.
Responses with no mention of the specific diagnosis(es)
they disagreed with were excluded. Braun and Clarke’s
(2006) methodology for thematic analysis was followed.
This enabled exploration of the research question in a data-
driven manner informed by participants’ perspective and
experience. S.K.A-Y. read through the responses to gain an
overall impression of the data set and noted any emerging
patterns. The responses were then reread and initial codes
were generated. Codes describing similar concepts were
grouped into preliminary themes and subthemes, which
were subsequently reviewed by A.E.R., L.B. and S.C. Dis-
cussion among the authors led to further refinement of the
Possibly autistic (n = 71) Non-autistic (n = 141)
40.1 (8.9), 20–57 42.1 (10.6), 20–60
11 (15.5%) 20 (14.2%)
60 (84.5%) 121 (85.8%)
35.8 (5.9) 17.5 (8.7)
 
Autistic Possibly autistic Non-autistic
(n = 208) (n = 71) (n = 141)
183 (88%) 57 (80.3%) 74 (52.5%)
25 (12.0%) 14 (19.7%) 67 (47.5%)
themes. Thereafter, related themes and subthemes were
consolidated, while those that were deemed to not mean-
ingfully address the research question were discarded.
Results
Rate of lifetime mental health diagnoses and
other conditions
There were significant between-group differences in the
proportions of participants reporting mental health or other
diagnoses, χ2(2) = 57.5, p < 0.001, V = 0.370 (see Table 2).
Autistic and possibly autistic participants were more likely
to report receiving mental health or other diagnoses com-
pared to non-autistic participants (p < 0.05). There was no
significant difference between autistic and possibly autis-
tic participants (p < 0.05).
Thereafter, we compared the proportion of individuals
reporting specific mental health diagnosis between groups
(see Table 3 for omnibus test statistics and pairwise com-
parisons). Types of diagnoses that participants selected or
mentioned were grouped for analysis based on Diagnostic
and Statistical Manual of Mental Disorders (5th ed.;
DSM-5 classifications; APA, 2013). There were no signifi-
cant between-group differences in proportion of individu-
als who reported diagnosis of Schizophrenia and Other
Psychotic Disorders, Bipolar and Related Disorders,
Dissociative Disorders, Gender Dysphoria and Substance
Related and Addictive Disorders. However, there were sig-
nificant between-group differences in the proportion of
individuals reporting diagnoses of Depressive Disorders,
Anxiety Disorders, Obsessive Compulsive and Related
Disorders, Trauma and Stressor Related Disorder, Feeding
and Eating Disorders and Personality Disorders, with con-
sistently greater number of participants in the autistic and/
or possibly autistic groups reporting these diagnoses.
Au-Yeung et al. 5

Table 3.  Number of participants who reported mental health diagnoses.

Autistic Possibly autistic Non-autistic Omnibus test Pairwise

(n = 208) (n = 71) (n = 141) comparisons
Depressive Disorders 161 (77.4%) 52 (73.2%) 60 (42.6%) χ2(2) = 47.4, p < 0.001, ASC > NASC
V = 0.336 PASC > NASC
ASC = PASC
Anxiety Disorders 148 (71.2%) 45 (63.4%) 48 (34%) χ2(2) = 48.6 p < 0.001, ASC > NASC
V = 0.340 PASC > NASC
ASC = PASC
Obsessive-Compulsive 40 (19.2%) 8 (11.3%) 3 (2.1%) χ2(2) = 23.1, p < 0.001, ASC > NASC
and Related Disorders V = 0.235 PASC > NASC
ASC = PASC
Personality Disorders 25 (12.0%) 8 (11.3%) 6 (4.3%) χ2(2) = 6.41, p = 0.041, ASC > NASC
V = 0.124 ASC = PASC
PASC = NASC
Feeding and Eating 21 (10.1%) 8 (11.3%) 4 (2.8%) χ2(2) = 7.49, p = 0.024, ASC > NASC
Disorders V = 0.134 PASC > NASC
ASC = PASC
Bipolar and Related 12 (5.8%) 5 (7%) 2 (1.4%) Fisher’s exact test = 5.49,  
Disorders p = 0.051, V = 0.108
Schizophrenia 7 (3.4%) 3 (4.2%) 1 (0.7%) Fisher’s exact test = 3.49,  
Spectrum and Other p = 0.149, V = 0.087
Psychotic Disorders
Trauma and Stressor 6 (2.9%) 6 (8.5%) 1 (0.7%) Fisher’s exact test = 8.19, PASC > NASC
Related Disorders p = 0.012, V = 0.150 ASC = NASC
ASC = PASC
Gender Dysphoria 2 (1%) 0 (0%) 1 (0.7%) Fisher’s exact test = 0.505,  
p = 1.00, V = 0.041
Dissociative Disorders 0 (0%) 1 (1.4%) 0 (0%) Fisher’s exact test = 3.46,  
p = 0.169, V = 0.108
Substance Related and 0 (0%) 1 (1.4%) 0 (0%) Fisher’s exact test = 3.46,  
Addictive Disorders p = 0.169, V = 0.108

ASC: autistic group; NASC = non-autistic group; PASC = possibly autistic group. More detailed analyses were carried out on the subtypes of
personality disorders, but no significant between-group differences were found (see Supplemental Materials A).

Table 4.  Number of participants who show agreement or disagreement with their mental health diagnosis.

‘Do you agree with your Autistic Possibly autistic Non-autistic

mental health diagnosis/es?’ (n = 182) (n = 57) (n = 74)
Yes 106 (58.2%) 26 (45.6%) 64 (86.5%)
No 12 (6.6%) 2 (3.5%) 3 (4.1%)
Not entirely 64 (35.2%) 29 (50.9%) 7 (9.5%)

Disagreement with mental health diagnosis
There were significant between-group differences in the
proportion of individuals who reported disagreement with
their mental health diagnoses as denoted in response to an
open-ended question on this topic, Fisher’s exact test = 31.8,
p < 0.001, V = 0.218 (see Table 4). Two of the post-hoc
pairwise comparisons were significant: these were the 2
(non-autistic vs autistic) × 2 (Yes vs Not entirely) compari-
son and the 2 (non-autistic vs possibly autistic) × 2 (Yes vs
Not entirely) comparison (both ps < 0.001). These showed
that a higher proportion of autistic and possibly autistic par-
ticipants did not entirely agree with their mental health
diagnosis compared to non-autistic participants, while a
greater proportion of non-autistic participants agreed with
their mental health diagnosis compared to the other two
groups. None of the other seven comparisons reached sig-
nificance (ps > 0.05). Proportions of participants who disa-
greed with specific type of mental health diagnoses were
compared between groups, but no significant differences
were found (see Supplemental Materials B).
6 Autism 00(0)
Thematic analysis
This section explored the reasons underlying disagreement
with mental health diagnosis(es). Data for the autistic and
possibly autistic participants were analysed together,
because our quantitative analysis showed that both groups
were significantly more likely to report disagreement with
their mental health diagnoses compared to the non-autistic
group (see Table 4). Therefore, it is likely that participants
in the possibly autistic group have shared experiences with
those in the autistic group. Spelling and grammatical errors
were corrected in participants’ extracts; these are denoted
in square brackets in the quotes. Each quote is followed
by a participant number and their diagnostic status
(ASC = autistic, PASC = possibly autistic). Data from the
non-autistic group were not analysed as very few of these
participants showed disagreement and responded to the
open-ended question. Two major themes emerged from the
thematic analysis: ‘Problems with Mental Health
Diagnosis’ and ‘Clinical Barriers’.
Problems with mental health diagnosis.  This theme consists
of two sub-themes: ‘ASC characteristics confused with
symptoms of a mental health condition’ and ‘mental health
difficulties perceived to be resultant of ASC’.
ASC characteristics confused with symptoms of a mental
health condition. The first most common reason for par-
ticipants’ disagreement with their mental health diagnoses
was that they felt their ASC characteristics were mistaken
as symptoms of a mental health condition. For example, in
the following extract, an autistic participant explains how
their difficulties with emotional regulation and self-inju-
rious behaviours may have been misinterpreted as symp-
toms of a borderline personality disorder (BPD):2
… [it’s] just aspergers with me, the out coming behaviour
might look or even be the same sometimes ([self-harm],
getting [emotional] etc) but the causes are completely
different- [I] learnt to hide my emotions and feelings to
survive school and home without being hurt so only [got]
visibly upset in the last moment when [it became] unbearable-
just because they cannot read my face [doesn’t] mean [I’m]
not having those emotions before … (153-ASC)
The same participant also reported that their lack of eye
contact was mistaken as evidence of schizophrenia rather
than social communication difficulty in ASC:
… I was given anti-psychotics for my behaviour … they think
because [I don’t] look at them and [I] am nervous of talking
then [I] am ‘guarded’ or have ‘flat affect’ … (153-ASC)
One possibly autistic participant did not ‘entirely agree’
with their Trauma and Stressor Related Disorder diagnosis
because they felt their difficulties with resistance to change
could be better explained by ASC, and specifically, the
category of ‘restricted, repetitive patterns of behaviour,
interests, or activities’ (APA, 2013):
… I think that ties in with possibly having an ASC as not
coping well with change is a characteristic of that.
(445-PASC)
Similarly, another possibly autistic participant who
received a bipolar disorder diagnosis felt their intense
interest had been mistaken as a sign of mania:
… I get mood swings, but I believe that is more due to the
Autism than manic depression. I am subdued when I am
uncomfortable and what may look manic when I am doing
something I like … (238-PASC)
Some participants spoke about sensory sensitivities that
they felt were mistaken as eating disorders or psychotic
symptoms:
My issues with food were regarding the texture and the act of
eating. (452-PASC)
I stopped going to the school canteen as a teenager and lost a
lot of weight because I couldn’t stand the social/sensory
environment. (2-ASC)
I am also not psychotic – I have [enhanced] sensory perception.
(28-ASC)
In addition to having autism characteristics miscatego-
rised by healthcare professionals, some participants
thought clues that were indicative of ASC might have been
missed. One autistic participant diagnosed with personal-
ity disorder reported:
… autism runs so strongly in my family (all four children
have AS/autism, and I think my father probably has Asperger
Syndrome). (175-ASC)
Likewise, a possibly autistic participant who was diag-
nosed with OCD reported,
… I suspect that is really ASC, but it was the 80s and I am
female. I have never been able to deal with much of anything
in the morning, and apparently shoe tying before school was
the last straw when little. (741-PASC)
These extracts highlight that the importance of family
history and behavioural patterns in early childhood in
making differential diagnosis between mental health con-
ditions and ASC.
Mental health difficulties perceived to be resultant of
ASC.  The second most common reason for participants’
disagreement with their mental health diagnoses was
because they felt that their mental health conditions were
resultant of the difficulties they experienced as part of
Au-Yeung et al. 7
living with ASC. These feelings were especially preva-
lent in participants who received depression and anxiety
diagnoses (nine participants). One autistic participant
explained that their depression stemmed from difficulties
connecting with others:
… because I did not ever really fit in socially and I couldn’t
understand what I was doing wrong, why people made friends
but didn’t include me, why I could never work out when to
join in a conversation in a group situation etc, I was depressed
because I thought I must be a very bad [unlovable] person.
Now I know I definitely have Aspergers I understand why I
don’t often connect with others properly. (120-ASC)
Another autistic participant reported that their depres-
sion and anxiety were a result of difficulties dealing with
social demands:
Depression and anxiety were a reaction to stress associated
with having autistic spectrum difficulties. I am not anxious
and depressed when I am in a suitable environment [emphasis
added]. (468-ASC)
Daily living can be exhausting for autistic individuals,
which could manifest as symptoms of depression. This
was experienced by the following participant who disa-
greed with her depression diagnosis:
… I live in a society that is not a suitable environment in
which I can thrive, as I am autistic. It is hard to access
accommodations. People are not kind and ask ignorant and
hurtful questions, and make harmful assumptions when I
plead with them to be accommodating as I am autistic. This is
quite taxing, and I often feel very tired. (480-ASC)
An autistic participant felt that due to their autism not
being understood, healthcare professionals tended to focus
on the co-occurring mental health condition – a more useful
treatment would be to address their psycho-social needs:
… My underlying ASC is hard to explain [to] the GP in that I
have expressed that I feel my depression is only a result of
my, kind of, ‘untreated Autism’ in effect. Meaning, that if I
can find help through relevant counselling, connecting with
others, to accept and understand the autistic side of me and
discover what my real needs are then I have a chance at
learning a new vocabulary to express those needs and this will
greatly reduce the resultant mental health issues that [recur],
particularly depression. (178-ASC)
This kind of support, however, may be difficult to
access – the second theme explores why this is the case.
Clinical barriers.  This theme describes the clinical barriers
participants faced in seeking accurate diagnosis and sup-
port. These barriers were related to the participants’ inter-
actions and experiences with healthcare professionals,
who they felt either:
… didn’t understand me. (1-ASC)
Or lacked an understanding of ASC:
… [the] health professionals who diagnosed BPD aren’t
familiar enough with the symptoms of a female with AS …
Now it feels like I’ve got a BPD diagnosis and they refuse to
look at anything else. (196-ASC)
Other participants felt some healthcare professionals
did not communicate with them. The following partici-
pants reported not having appropriate discussions regard-
ing their diagnoses, and not being listened to, or given
opportunities to raise concerns:
I never had a proper assessment or discussion for my diagnosis
of EUPD2 and nobody told me, it just appeared on my medical
records. (586-PASC)
… I was misdiagnosed but no one would listen to me and I
was told I was lying when I tried to explain. (2-ASC)
… After obtaining medical notes, it is clear that my account of
my experiences were being misunderstood and I was not
given opportunities to clarify as information was withheld
from me … (8-ASC)
The consequence of a lack of autism awareness and
communication was that it hindered participants receiving
adequate support and having their needs met. The follow-
ing autistic participants reported dissatisfaction with medi-
cation being the focus of their treatment:
… I feel my GP is focusing purely on the depression and once
he has a ‘result’ through the anti-depressants, then I will be
deemed fully well again. Although I have taken the anti-
depressants for 8 weeks now, I still struggle with dark
thoughts, feelings of wanting to end it all and find life poses
the same difficulties to me when it comes to functioning.
(178-ASC)
… my GP tries to keep throwing antidepressants at me but [I]
think [I] just need to understand myself better [and] get other
conditions diagnosed. (468-ASC)
Similarly, a possibly autistic participant talked about
comparable experience with regard to therapeutic inter-
ventions received, with the focus on the treating the men-
tal health condition overshadowing their undiagnosed
autism:
… I’ve had therapy with MIND3 and [counselling] sessions,
although they helped at the time, no one got to the bottom of
what was wrong. (24-PASC)
The following participant’s experience is a fitting illus-
tration of this theme and the difficulties participants face in
seeking accurate diagnosis and support. This participant
was misdiagnosed with bipolar and prescribed various
8 Autism 00(0)
medications over many years to manage a condition they
did not have:
… I had a formal diagnosis of bipolar (II) condition for around
17 years. I actually satisfy the DSM criteria for this, but have
only [ever] been hypomanic once and that was antidepressant
induced. I don’t benefit from mood stabilisers, antipsychotics
or most antidepressants, and things like stress, working long
days and staying up late don’t induce mania. I have tried lots
of different medications over the years, but for most of my
[years] with a bipolar diagnosis, I worked in developing
countries, didn’t take any [psychotropic medications] and
remained relatively mentally well … Hence, once my autism
had been diagnosed, I was able to get a consultant psychiatrist
to say that she didn’t think I’d ever had bipolar and it has been
revoked as a current diagnosis. (339-ASC)
Discussion
This study explored the experience of mental health diag-
nosis comparing non-autistic, autistic and possibly autistic
people and addressed the following research questions: (1)
Are autistic individuals more likely to report receiving
mental health diagnosis(es) and if so, what kind of diagno-
ses are they more likely to receive? (2) Are autistic indi-
viduals more likely to disagree with these diagnoses? (3)
What were the reasons behind any disagreement? We
found that autistic and possibly autistic adults reported
receiving more mental health diagnoses compared to non-
autistic adults. Specifically, significant differences were
found for Depressive Disorders, Anxiety Disorders,
Obsessive Compulsive and Related Disorders, Trauma and
Stress Related Disorders, Feeding and Eating Disorders
and Personality Disorders. Autistic and possibly autistic
adults were also significantly less likely to agree with their
mental health diagnoses than non-autistic adults. There
were two major reasons for their disagreement: (1) ASC
characteristics were confused with symptoms of a mental
health condition and (2) mental health difficulties per-
ceived to be resultant of ASC. Autistic and possibly autis-
tic individuals spoke about the clinical barriers that hinder
accurate diagnosis and support, which include healthcare
professionals’ lack of awareness and understanding of
autism, poor communication between autistic adults and
healthcare professionals and treatment not being suited to
their needs.
Our findings are consistent with previous research doc-
umenting that autistic people (both diagnosed and undiag-
nosed) experience greater rates of co-occurring mental
health conditions compared to the general population
(Lever and Geurts, 2016). There may be several reasons as
to why autistic people are more vulnerable to developing
mental health conditions. For example, Haruvi-Lamdan
et al. (2017) suggested that autistic people are more likely
to experience stressful or traumatic life events such as
being bullied or ostracised; cognitive features (e.g. having
rigid thinking and rumination) may also affect how stress-
ful events are processed; additionally, sensory sensitivities
and sudden changes in routines and environment are a
source of stress for autistic individuals over and above
what is experienced by non-autistic individuals. Although
the effect of these factors on autistic people’s mental health
is not yet fully understood, it seems clear that these are
significant and widespread issues within the autistic
population.
Our findings showed that only half of the participants in
both the autistic (58.2%) and possibly autistic (45.5%)
groups agreed with the mental health diagnosis(es) they
received, compared to over 86.5% in the non-autistic
group. Qualitative responses from autistic and possibly
autistic participants indicated that similarities in symptom
presentation between ASC and a range of mental health
conditions could be the source of these perceived misdiag-
noses. Crane et al. (2018) conducted a survey and inter-
views exploring the mental health experience of young
autistic people (16–25 years). They found that young autis-
tic people perceive themselves as different from non-autis-
tic people in many domains (e.g. that they are more likely
to feel under strain or unhappy) even when they are not
experiencing a mental health problem. This has implica-
tions for the development of diagnostic tools, because it
could be the case that measures developed with the general
population in mind may therefore fail to tease apart
whether an autistic person is experiencing psychological
distress that is over and above their typical behavioural
presentation. One of the current research priorities of our
group is to develop validated diagnostic tools to accurately
assess mental health conditions in the autistic population
(Cassidy et al., 2018a, 2018b).
Another contributing factor to the participants’ per-
ceived misdiagnosis was the lack of awareness and under-
standing of ASC among healthcare professionals. This
study highlights the importance of autism awareness train-
ing for health and social care professionals, as outlined by
statutory guidance (Department of Health, March 2015).
Such training would arm professionals with the confidence
to communicate with, understand and support autistic peo-
ple. Mental health professionals should be trained to
understand the cognitive differences between autistic and
neurotypical people, which will help them modify thera-
peutic interventions so they are effective for autistic peo-
ple. In addition, more research is needed to test the
effectiveness of modified therapeutic interventions and
medications for various mental health conditions for the
autistic population.
An important topic arose from our qualitative analysis
relating to diagnostic overshadowing. Our participants
reported that they felt healthcare professionals gave prior-
ity to treating mental health conditions and thus leaving
the underlying autism unaddressed and other psycho-
social needs unmet. Similarly, in the study by Crane et al.
Au-Yeung et al. 9
(2018), interviews with young autistic people highlighted
numerous barriers for support. These included services not
being tailored to their needs in relation to their autism,
diagnostic and mental health services being ‘disjointed’,
and worse, available support becoming even scarcer when
they transitioned to adulthood. Given that many autistic
adults received their diagnosis in adulthood, there is a real
need to develop adequate post-diagnostic support systems.
Post-diagnostic support could include: psycho-social edu-
cation; provision of opportunities for autistic people to
connect with others through local support networks; advo-
cacy services to support autistic people to express their
needs; and signposting to local resources. Mainstream ser-
vices such as adult mental health services may also benefit
from the support and consultation of a specialised ASC
service in the local area (see National Institute for Health
and Care Excellence (NICE), 2014, for the Bristol Model).
Strengths and limitations
A key strength and novel aspect of this study was the par-
ticipatory approach, where autistic adults prioritised the
study topic and questions utilised in the research. This
ensured that the study was relevant, important and clear to
autistic people and thus more likely to be useful to their
daily lives – a feature lacking in much previous autism
research (Fletcher-Watson et al., 2018; Pellicano et al.,
2014). Furthermore, we included possibly autistic people
in the study who had not yet been formerly diagnosed, but
strongly identified as autistic, as suggested by our autistic
steering group. This inclusion allowed us to explore
whether yet-to-be-diagnosed autistic adults shared similar
experiences with those with a clinically confirmed diagno-
sis of ASC, which was important given many autistic
adults remained undiagnosed or received a diagnosis much
later in life (Lai and Baron-Cohen, 2015). Interestingly,
the results from the confirmed and possibly autistic groups
were very similar. The mean AQ score of the possibly
autistic group (35.8) was well above the recommended
cut-off (32+) for clinically significant level of autistic
traits, and similar rates of mental health diagnoses and
level of disagreement indicated shared experiences
between these groups. However, we do acknowledge that
there is no way of knowing whether participants in the
possibly autistic group truly met the diagnostic criteria of
ASC without a comprehensive clinical assessment.
We found lifetime rate of mental health and other diag-
noses to be 88% for the autistic group and 52.5% for the
non-autistic group. Previous work by Lever and Geurts
(2016) reported similar lifetime rates of psychiatric condi-
tions in their autistic sample (79%) and non-autistic group
(48.8%). This shows that that the self-reported rates of
mental health diagnoses in our study are consistent with
previously reported rates using standardised instruments
for diagnosing mental health disorders. However, we are
unable to verify whether the self-reported rates of disa-
greement consistent with actual rate of misdiagnosis due to
the lack of previous research in this area. We are also una-
ble to conclude whether autistic and possibly autistic indi-
viduals were more likely to disagree with specific mental
health diagnoses that they received because of the small
sample size for the disagreement data; therefore, these
comparisons are likely to lack power and should be inter-
preted with caution (see Supplemental Materials B).
Finally, the self-selecting nature of our sample meant
that it was not possible for us to control the variability in
age and gender. How might these confounding variables
have influenced our findings? First, we might expect that
participant groups with higher age to have a higher rate of
lifetime mental health diagnosis, however, this was not the
case; our autistic group reported significantly more life-
time mental health diagnoses despite being statistically
younger than the non-autistic group. Second, we might
expect that the participant group with greater proportion of
females to have higher rates of lifetime mental health diag-
nosis based on findings previously reported for females, in
particular, for mood and anxiety disorders (Riecher-
Rössler, 2016). This is again not the case in our sample;
our autistic group reported significantly higher rates of
mental health diagnosis despite having a smaller propor-
tion of females. While these confounding variables do not
seem to have affected the findings in the predicted manner,
we acknowledge that the confounding variables still pose a
limitation to the study and may affect the generalisability
of our findings.
Acknowledgements
The authors sincerely thank members of the Coventry Autism
Steering Group, who assisted the researchers in designing and
advertising the study, and Paula Smith, database manager at the
Autism Research Centre, University of Cambridge for her assis-
tance with contacting participants registered in the Cambridge
Autism Research Database. They also thank Dr Jacqui Rodgers,
Professor Clare Wood, David Mason, Professor Helen
McConachie, Professor Rory O’Connor, Professor David Mosse,
Dr Carrie Allison and Dr Rebecca Kenny for valuable discus-
sions and also thank everyone for taking part in the study. They
appreciate that this is a difficult topic to think and talk about and
appreciate their support in increasing understanding and preven-
tion of mental health problems, self-injury and suicide.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with
respect to the research, authorship and/or publication of this
article.
Funding
The author(s) disclosed receipt of the following financial support
for the research, authorship and/or publication of this article:
This work was supported by the Economic and Social Research
Council (grant number ES/N000501/2), Autistica (grant number
10 Autism 00(0)
7247) and a research pump prime from Coventry University
(received by S.C.). S.B-C. was supported by the Autism Research
Trust, the MRC, the National Institute for Health Research
(NIHR) Collaboration for Leadership in Applied Health Research
and Care East of England at Cambridgeshire and Peterborough
NHS Foundation Trust and the Innovative Medicines Initiative
Joint Undertaking under grant agreement no. 115300, resources
of which are composed of financial contribution from the
European Union’s Seventh Framework Programme (FP7/2007-
2013) and EFPIA companies’ kind contribution. The views
expressed are those of the authors and not necessarily those of the
NHS, the NIHR or the Department of Health. The funders had no
role in the data collection, analysis or interpretation or any aspect
pertinent to the study or publication. The corresponding author
has full access to the data in the study and had final responsibility
for the decision to submit for publication.
Notes
1. We use the term condition, rather than disorder, in recogni-
tion of the strengths and difficulties associated with autism
and in order to be less stigmatising. Recent research has also
found that the autistic community uses a range of terms to
describe autism, but the term ‘autistic person’ is most pre-
ferred (Kenny et al., 2016). Thus, we use the terms ASC,
autism and autistic synonymously throughout this article in
recognition of this.
2. The abbreviated terms for Borderline Personality Disorder
(BPD) and Emotionally Unstable Personality Disorder
(EUPD) were used interchangeably in participants’ extract;
these are synonyms for the same condition.
3. Mind is a registered mental health charity and company that
provides advice and support to anyone experiencing a men-
tal health problem across England and Wales.
Supplemental material
Supplemental material for this article is available online.
ORCID iDs
Sheena K Au-Yeung https://orcid.org/0000-0002-0516-4755
Ashley E Robertson https://orcid.org/0000-0001-9914-3199
Sarah Cassidy https://orcid.org/0000-0003-1982-3034
References
American Psychiatric Association (2013) Diagnostic and
Statistical Manual of Mental Disorders. 5th ed. Washington,
DC: American Psychiatric Association.
Baron-Cohen S, Wheelwright S, Skinner R, et al. (2001) The
Autism-Spectrum Quotient (AQ): evidence from Asperger
syndrome/high-functioning autism, males and females,
scientists and mathematicians. Journal of Autism and
Developmental Conditions 31(1): 5–17.
Braun V and Clarke V (2006) Using thematic analysis in psy-
chology. Qualitative Research in Psychology 3(2): 77–101.
Brugha TS, McManus S, Bankart J, et al. (2011) Epidemiology
of autism spectrum disorders in adults in the community in
England. Archives of General Psychiatry 68(5): 459–465.
Camm-Crosbie L, Bradley L, Shaw R, et al. (2018) ‘People like
me don’t get support’: autistic adults’ experiences of sup-
port and treatment for mental health difficulties, self-injury
and suicidality. Autism. Epub ahead of print 29 November.
doi: 10.1177/1362361318816053
Cassidy SA, Bradley L, Bowen E, et al. (2018a) Measurement
properties of tools used to assess depression in adults with
and without autism spectrum conditions: a systematic
review. Autism Research 11(5): 738–754.
Cassidy SA, Bradley L, Bowen E, et al. (2018b) Measurement
properties of tools used to assess suicidality in autistic and
general population adults: a systematic review. Clinical
Psychology Review 62: 56–70.
Cassidy SA, Bradley P, Robinson J, et al. (2014) Suicidal idea-
tion and suicide plans or attempts in adults with Asperger’s
syndrome attending a specialist diagnostic clinic: a clinical
cohort study. The Lancet Psychiatry 1(2): 142–147.
Cassidy SA, Bradley L, Shaw R, et al. (2018c) Risk markers for
suicidality in autistic adults. Molecular Autism 9: 42.
Cohen J (1988) Statistical Power and Analysis for the Behavioral
Sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum
Associates, Inc.
Crane L, Adams F, Harper G, et al. (2018) ‘Something needs to
change’: mental health experiences of young autistic adults
in England. Autism. Epub ahead of print 1 February. DOI:
10.1177/1362361318757048.
Department of Health (2015) Statutory guidance for local author-
ities and NHS organisations to support implementation of
the adult autism strategy. Available at: https://www.gov.
uk/government/uploads/system/uploads/attachment_data/
file/422338/autism-guidance.pdf
Dosseter DR (2007) ‘All that glitter is not gold’: misdiagnosis
of psychosis in pervasive developmental disorders – a case
series. Clinical Child Psychology and Psychiatry 12(4):
537–548.
Fletcher-Watson S, Adams J, Brook K, et al. (2018) Making the
future together: shaping autism research through meaning-
ful participation. Autism. Epub ahead of print 10 August.
DOI: 10.1177/1362361318786721.
Haruvi-Lamdan N, Horesh D and Golan O (2017) PTSD and
autism spectrum disorder: co-morbidity, gaps in research,
and potential shared mechanisms. Psychological Trauma:
Theory, Research, Practice and Policy 10: 290–299.
Hedley D, Uljarević M, Foley K-R, et al. (2018) Risk and protective
factors underlying depression and suicidal ideation in autism
spectrum disorder. Depression and Anxiety 35: 648–657.
Hirvikoski T, Mittendorfer-Rutz E, Boman M, et al. (2016)
Premature mortality in autism spectrum disorder. British
Journal of Psychiatry 208(3): 232–238.
Hoekstra RA, Bartels M, Verwejj CJ, et al. (2007) Heritability
of autistic traits in the general population. Archives of
Pediatrics and Adolescent Medicine 161(4): 372–377.
Jones L, Goddard L, Hill EL, et al. (2014) Experiences of receiv-
ing a diagnosis of autism spectrum disorder: a survey
of adults in the United Kingdom. Journal of Autism and
Developmental Disorders 44(2): 3033–3044.
Kenny L, Hattersley C, Molins B, et al. (2016) Which terms
should be used to describe autism? Perspectives from the
UK autism community. Autism 20(4): 442–462.
Lai M-C and Baron-Cohen S (2015) Identifying the lost genera-
tion of adults with autism spectrum conditions. The Lancet
Psychiatry 2(11): 1013–1027.
Au-Yeung et al. 11
Lever AG and Geurts HM (2016) Psychiatric co-occurring
symptoms and disorders in young, middle-aged, and older
adults with autism spectrum disorder. Journal of Autism and
Developmental Disorders 46(6): 1916–1930.
Luciano CC, Keller R, Politi P, et al. (2014) Misdiagnosis of high
function autism spectrum disorders in adults: an Italian case
series. Autism 4: 131.
Matson JL and Williams LW (2013) Differential diagnosis and
comorbidity: distinguishing autism from other mental health
issues. Neuropsychiatry 3(2): 233–243.
National Institute for Health and Care Excellence (2014)
Implementation pack: developing a multi-agency local
autism team. Putting NICE guidance into practice. Available
at: https://www.nice.org.uk/guidance/cg170/resources/the-
bristol-autism-spectrum-service-bass-pdf-436748082
Pellicano E, Dinsmore A and Charman T (2014) What should
autism research focus upon? Community views and priori-
ties from the United Kingdom. Autism 18(7): 756–770.
Pelton MK and Cassidy SA (2017) Are autistic traits associated
with suicidality? A test of the interpersonal-psychological
theory of suicide in a non-clinical young adult sample.
Autism Research 10(11): 1891–1904.
Riecher-Rössler A (2016) Sex and gender differences in mental
disorders. The Lancet Psychiatry 4(1): 8–9.
Rynkiewicz A, Schuller B, Marchi E, et al. (2016) An investi-
gation of the ‘female camouflage effect’ in autism using a
computerized ADOS-2 and test of sex/gender differences.
Molecular Autism 7: 10.
Stewart M, Barnard L, Pearson J, et al. (2006) Presentation of
depression in autism and Asperger syndrome: a review.
Autism 10: 103–116.
Van Schalkwyk GI, Peluso F, Qayyum Z, et al. (2015) Varieties
of misdiagnosis in ASD: an illustrative case series.
Journal of Autism and Developmental Disorders 45(4):
911–918.
Woodbury-Smith MR, Robinson J, Wheelwright S, et al. (2005)
Screening adults for Asperger syndrome using the AQ: a
preliminary study of its diagnostic validity in clinical prac-
tice. Journal of Autism and Developmental Disorders 35(3):
331–335.